The timely selection of the best treatment for patients with depression is critical to the goal of improving remission rates. Due to the biological heterogeneity and variable symptom presentation of depression, it is unlikely that a single clinical or biological marker can guide treatment selection. Rather, a biosignature developed from a systematic exploration of a group of clinical and biological markers is more likely to be successful. Two types of biosignatures are needed to achieve improved outcomes: 1) biosignatures to maximize the selection of optimal treatment for individual patients at the beginning of treatment (moderators) and 2) biosignatures to identify indicators of eventual outcomes early in treatment (mediators). This approach has great potential to personalize treatment and maximize the number of patients who can be treated to full remission with a given treatment. We propose a comparative effectiveness trial of three mechanistically distinct treatments for MDD (citalopram, bupropion, and cognitive behavioral therapy) in which we will assess a comprehensive array of carefully selected clinical (i.e. anxious depression, early life trauma, & gender) and biological (i.e. genetic, neuroimaging, serum, epigenetic & qEEG) moderators and mediators of outcome. Using innovative statistical approaches the identified moderators and mediators will then be used to develop a differential depression treatment response index (DTRI). The proposed study is a randomized two-stage trial (Stagel:12 wks; Stage2: 12 wks) design with 675 MDD patients (with a history of one adequate trial of an SSRI except citalopram) assigned to one of three treatment conditions (n=225 each). This two stage approach is similar to a Sequential Multiple Assignment Randomized Trial (SMART) design. This application brings together researchers with extensive experience in conducting large clinical trials and experts at the forefront of the neurobiology of depression, including: clinical trials (Trivedi, Fava, Schatzberg,Nierenberg, Shelton, Gaynes, Hollon), genetics (Smoller, Binder, McMahan, Perils), neuroimaging (Phillips, Sheline, Etkin, Pizzagalli, Buckner), qEEG (losifescu, Ellenbogen), neurotrophins/cytokines (Duman, Sanacora, Turck, Shelton), clinical predictors (Shelton, Hollon, Trivedi, Fava, Nierenberg, Goodman, Yehuda), neuroendocrine markers (Holsboer, Schatzberg, Shelton, Yehuda), epigenetics (Nestler, Yehuda),and cognitive behavior therapy (Hollon, Manber, Arnow). This team will also be guided by internationally known biomarker scientists (Holsboer, Schatzberg, Krystal, Charney, Goodman), as well as a highly qualified group of biostatisticians (Kraemer, Wisniewski, Schoenfeld).